Seliverstov

This article deals with co-algorithmic hybridization of two Particle Swarm Optimization (PSO) algorithms, using click-type and ring-type neighborhood topologies of particles. Each of these topologies has their disadvantages which can be eliminated by using co-algorithmic hybridization. The authors compared the efficiency of canonical PSO and its co-algorithmic modification and found optimal values of free parameters.

The article describes a parallel method of global unconditional optimization by particle swarm. Features of implementation in CUDA graphics processors are described. Study results of convergence, performance and scalability are given.

This paper presents review of the particle swarm optimization method (PSO) for a global optimization problem and its certain serial and parallel modifications.